* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download Martinez (2010) 1 Chapter 2 Week 3 Gredler (2009)
Prosocial behavior wikipedia , lookup
Cognitive science wikipedia , lookup
Social Bonding and Nurture Kinship wikipedia , lookup
Social psychology wikipedia , lookup
Conservation psychology wikipedia , lookup
Observational methods in psychology wikipedia , lookup
Educational psychology wikipedia , lookup
Symbolic behavior wikipedia , lookup
Abnormal psychology wikipedia , lookup
Thin-slicing wikipedia , lookup
Transtheoretical model wikipedia , lookup
Neuroeconomics wikipedia , lookup
Behavioral modernity wikipedia , lookup
Classical conditioning wikipedia , lookup
Organizational behavior wikipedia , lookup
Applied behavior analysis wikipedia , lookup
Insufficient justification wikipedia , lookup
Learning theory (education) wikipedia , lookup
Attribution (psychology) wikipedia , lookup
Theory of planned behavior wikipedia , lookup
Theory of reasoned action wikipedia , lookup
Adherence management coaching wikipedia , lookup
Descriptive psychology wikipedia , lookup
Verbal Behavior wikipedia , lookup
Behavior analysis of child development wikipedia , lookup
Social cognitive theory wikipedia , lookup
Psychological behaviorism wikipedia , lookup
Martinez (2010) 1 Chapter 2 Week 3 Gredler (2009) - Review Ch. 1, 2, 4, 5, 8, & 9 Ormrod (2008) - Review Ch. 3, 4, 7, & 11 Martinez (2010) - Review Ch. 2, 5, & 7 Learning and Cognition. The Design of the Mind Chapter 2: Behaviorism ISBN: 9780205507245 Author: Michael E. Martinez copyright © 2010 Pearson Education Behaviorism One particular theoretical framework predominates in this book—the framework of cognitive psychology. From the cognitive perspective, the human mind is appreciated as a personal entity capable of reasoning, solving problems, and learning complex information. The mind also feels emotions and has intentions; it has goals and plans. Almost every chapter of this book holds to the cognitive point of view because such a framework provides the educator with a rich set of concepts and models that can guide the design of learning experiences. Martinez (2010) 2 This chapter, however, begins the study of learning through the framework of behaviorism, the historical precursor of modern cognitive psychology. Many behaviorist ideas have strongly influenced education and continue to be used by teachers to this day. Parenting, as well as education, has also been profoundly influenced by behavioristic principles. Even the daily interactions of adults have been touched by the language and concepts taken directly from behaviorist theory. Why Study Behavior? Behaviorism is the study of learning in humans and animals as understood through the analysis of behavior rather than thoughts or feelings. It is forbidden, at least in strict forms of behaviorism, to account for learning (understood by behaviorists as changes in behavior) by invoking the mind or mentalistic concepts (Ringen, 1999). Accounts of behavior that invoke the organism’s goals, wishes, beliefs, choices, will, or hopes are considered to be outside the explanatory norms of behaviorism. So what counts as an acceptable explanation of behavior? The behaviorist is interested only in what can be observed directly. Legitimate explanations for behavior are therefore restricted to what can be observed objectively, such as what Martinez (2010) 3 happens to the organism before or after a particular behavioral act. Advantages and Disadvantages of Behaviorism As we approach behaviorism, we need to appreciate its advantages as well as its limitations. One advantage of behaviorism over other approaches to understanding learning can be stated succinctly: By focusing strictly on behavior and on objective explanations for behavior, the methodology of behaviorism appears to be scientific. One potential problem with cognitive explanations of behavior is that such concepts as knowledge, goals, and memory are not directly observable. They must be inferred from what people say and do, leading to subjective and possibly incorrect interpretations. But when the discussion is restricted to what the person or animal actually does, agreement is much more likely. The early behaviorists, especially John B. Watson (who we will consider later in this chapter), were acutely aware of the methodological advantages of behaviorism as a scientific approach to understanding humans and animals. Even though there are distinct advantages to approaching human learning and behavior from the viewpoint of behaviorism, there are also definite Martinez (2010) 4 disadvantages. Consider what it would be like for a teacher never to use mental language—words such as remember, think, reason, decide, believe, and so on. Teachers tend to rely on language that is essentially cognitive. Not only teachers, but also everyday people in everyday situations use the language of the mind. Such language is useful. Moreover, theories that accept mental language and place the mind centrally in accounts of education can enrich the learning process substantially. Throughout this book, we will benefit greatly from cognitive theory as we approach the project of education. Nevertheless, our focus in this chapter is on behaviorism—a more conceptually restricted approach that focuses on what is objectively observable, and on explanations that do not appeal to the mind or to the mind’s presumed activity. Classical Conditioning The title of this section might provoke two questions: What does classical mean, and what does conditioning mean? To a behaviorist, conditioning is a synonym for learning, which, in turn, refers to a change in behavior. This way of defining learning seems narrow, but because behaviorists concentrate on external events (e.g., rewards and punishments), not on internal ideas (e.g., knowledge or intentions), the definition fits. Martinez (2010) 5 Conditioning, then, refers to an enduring change in behavior that results from events in the organism’s environment. What then is the meaning of classical in classical conditioning? To answer, it is important to understand that within behaviorism there are two basic approaches to conditioning: classical conditioning and operant conditioning. Classical conditioning is primarily associated with the Russian physiologist Ivan Pavlov; operant conditioning is primarily associated with the American psychologist B. F. Skinner. Classical and operant conditioning are quite different, and each is very important. Both are relevant to education. Ivan Pavlov’s Discovery Let’s start our study of behaviorism by considering a major discovery by the Russian scientist, Ivan Pavlov. Psychologists sometimes forget that Pavlov was already a famous scientist before he made any contributions to psychology. He was not merely famous; in 1904 he had won the highest achievement possible in the career of a scientist, the Nobel Prize, for discoveries in the physiology of digestion. Pavlov’s later discoveries in behaviorism (1927), and in particular his discovery of classical conditioning, occurred after he had proven himself as an eminent scientist in the field of physiology. Martinez (2010) 6 The breakthrough event that prompted the theory of classical conditioning took place in Pavlov’s laboratory, where experiments on physiology used dogs as subjects. In the late 1890s, Pavlov’s research initially focused on the chemistry of dog saliva. The volume of dog spit was carefully calibrated by precisely marked receptacles. Pavlov’s lab apparatus was also designed to deliver precise quantities of food to the dog’s mouth through a tube (Figure 2.1). Figure 2.1 Pavlov’s Dogs Even a neutral stimulus could prompt salivation. [Allyn & Bacon] There was a natural connection between the food delivered to the dog through a tube, and the amount of saliva collected from the dog by a different tube. Now consider parallels when one prompted the other. But something else did surprise Pavlov: The dog sometimes salivated before it was given the food. Pavlov wondered what besides food could trigger salivation in the dog. His best guess was that the dog responded to the sound of footsteps of his lab assistant. Martinez (2010) 7 Consider what it means for the sound of footsteps to prompt salivation. The sounds of someone walking would not ordinarily cause a dog to salivate. At some point, a new connection— footsteps and drooling—had to be learned, or in the language of behaviorists, it had to be conditioned. In this example of conditioning (learning) we can easily identify the cause (the sound of footsteps) and the effect (salivation). These are called, respectively, the stimulus and response. Putting this all together, we can express the connection between stimulus and response as follows: Footsteps → Dog Salivation (Conditioned Stimulus) (Conditioned Response) We know that the connection between food and salivation is quite natural. Saliva is produced in the mouth to aid in the digestion of food. The production of saliva does not need to be learned. So we can say that in the case of food there is an unconditioned stimulus and unconditioned response, denoted as follows: Food → Dog Salivation (Unconditioned Stimulus) (Unconditioned Response) One stimulus-response connection (involving footsteps) is learned, and the second (involving Martinez (2010) food) is not. One is conditioned; the other is unconditioned. The conditioned salivation was interesting to Pavlov because it was prompted by a totally neutral stimulus. Nothing about the sound of footsteps should make a human or a dog salivate. The sound of shoes on a floor is neutral with regard to saliva production. Pavlov, wondering if other neutral stimuli could evoke a salivation response, found that the sounds of whistles, metronomes, and tuning forks likewise evoked salivation if the sounds regularly preceded the administration of food. At this point in his research, Pavlov was on his way to forming a generalization that any neutral stimulus could be made to evoke a natural response through conditioning. Let’s not forget the nature of the response— salivation—is not sophisticated behavior; it’s primal. That is why Pavlov called the learned behavior in his studies conditioned reflexes. As behavior, salivation is about as instinctual as it gets. This is key to understanding classical conditioning. It’s not about sophisticated behavior. It’s about simple, very primitive behavior—and about what causes that behavior. Classical conditioning explains how a neutral (conditioned) 8 Martinez (2010) 9 stimulus, through learning, can evoke the same response as a natural (unconditioned) stimulus. One way of thinking about classical conditioning is that it entails stimulus substitution, where the conditioned stimulus substitutes for the unconditioned stimulus. For this to happen, the conditioned stimulus must become associated with the unconditioned stimulus—just as footsteps were associated with the dog food. Through research, Pavlov discovered that certain aspects of that association were important. For example, the conditioned stimulus has to occur before the unconditioned stimulus, rather than afterward. And the association between the two is strongest if the conditioned stimulus occurs just an instant before the unconditioned stimulus. A separation in time of more than a few seconds weakens the association and produces a weaker response. Why Is Classical Conditioning Important? All of this would be pretty uninteresting if the conditioned response was only salivation. But many other forms of primitive behavior are quite important to the survival of both humans and animals. Consider fear, for example. Some kinds of stimuli naturally evoke a fear response. Snakes often evoke fear not only in humans but also in higher primates such as gorillas and chimpanzees. Martinez (2010) 10 The fear of snakes is apparently unlearned; somehow it’s hard-wired into the brains of some species. Whenever a stimulus causes fear, it produces a physiological state that makes the heart pound faster. Blood is shunted to muscles to get the body ready to move aggressively toward or protectively away from the object of fear. This physiological state is sometimes called the fight or flight response. Experiences that naturally evoke fear include loud noises, angry faces, and threatening gestures. These are not neutral stimuli. But classical conditioning teaches that a neutral stimulus can evoke the same primitive behavior (such as a fear response) as a natural stimulus. In other words, we can learn to fear a neutral stimulus if, through our experiences, that fear has been conditioned over time. This means that negative emotions can, through conditioning, become associated with any stimulus. Perhaps you now see how classical conditioning explains how we can acquire strong emotional reactions to objects, events, and people that are initially neutral. When we speak of primitive behavior, such as salivation, we do not mean unimportant behavior. Negative emotional reactions, such as fear or feelings of revulsion, are highly significant. They Martinez (2010) 11 influence how we live our lives, and in the education arena they influence what and how we learn, as well as the decisions we make along our educational pathways. At issue here are not only behavioral responses that involve negative emotions. Positive emotions, such as interest and feelings of pleasure, are also prompted by certain natural stimuli (such as a teacher’s smiling face) and, by extension through conditioning, to all manner of neutral stimuli (such as books and classrooms). Early Behaviorism in America Although the Russian scientist Ivan Pavlov played a crucial role in discovering classical conditioning, behaviorism was mainly a phenomenon of American psychology. In fact, behaviorism dominated American psychology for a half-century, from about 1920 to about 1970. John B. Watson: Psychology as a Science If any one person is to be credited with launching this era on the North American continent, it is John B. Watson. In a 1913 article entitled “Psychology as the Behaviorist Sees It,” Watson argued vigorously that the only appropriate psychology is behaviorism. The basis for this claim is the arguably scientific quality of the behaviorist’s Martinez (2010) 12 methodology and theory. Behaviorists insisted on objective observations and measurement. Their theories, likewise, permitted only concepts that could be directly observable—that is, behavior— and shunned every idea and explanation that could not be observed directly and objectively, including mentalistic concepts. For Watson, these methodological and theoretical restrictions were essential if psychology was to progress as a scientific enterprise. Watson was also significant for showing how principles of conditioning applied to humans. Ivan Pavlov’s earlier work with animals showed that through stimulus association a neutral object could evoke a simple behavior. Watson found that conditioning also applied to people. In his most famous research, Watson conditioned a baby, Little Albert, to fear certain neutral objects. Watson began his experiment by showing that Little Albert had no natural fear of a white rat. When the rat appeared, Little Albert showed interest rather than fear. All that changed when Watson began to pair the appearance of the white rat with loud noises. When the rat appeared, Watson would bang on a loud pipe, causing Little Albert to cry out in fear. It is easy to predict that the fear reaction prompted by the jarring sound of banging pipes quickly became associated with the white rat. Martinez (2010) 13 Soon, the appearance of the rat alone, absent loud noises, could evoke a strong fear reaction in Little Albert. In Pavlovian fashion, the role played by the unconditioned stimulus (a loud noise) was taken up by the conditioned stimulus (the rat). Watson’s simple conditioning experiment vividly demonstrated behaviorism’s significance for understanding human learning and behavior. If fear reactions could be prompted by one neutral object, why not other objects? Indeed, the next phase of the Little Albert experiment showed that the conditioned response (fear) transferred to other neutral objects—even without additional training. All that was needed was to present stimuli that looked something like a white rat. When Watson showed fluffy white toys to Little Albert, the boy displayed the same fear reaction prompted by the white rat. Even a Santa Claus beard made Little Albert cry out in fear. The function of the original conditioned stimulus, the rat, had expanded to other, somewhat similar, stimuli. This expansion beyond the original stimulus, called stimulus generalization, is common in behavioristic conditioning. But stimulus generalization extends only so far. If a stimulus is too different from the original, the learned reactions will not be evoked. Presumably, Little Albert would not be afraid of a red fluffy toy. This Martinez (2010) 14 counterpart process is stimulus discrimination, in which a person or animal perceives differences between two stimuli such that one evokes a learned response and the other does not. John Watson was highly impressed with the power he had to shape the behavior patterns of Little Albert and others. To Watson, behaviorism was not simply the proper way to study human behavior, it was also a powerful means to alter human behavior. Watson was not thinking here of artificial kinds of conditioning such as the one he set up in his laboratory. He believed that every person is conditioned by the events of a unique life path. From infancy to old age, our behavior is learned through specific experiences, just as Little Albert’s was. The insight led Watson to conclude that human behavior is highly malleable—that any baby could be molded into any kind of adult, given the right training. This conviction led Watson (1924, p. 82) to make a stunning claim for which he became famous: Give me a dozen healthy infants, well-formed, and my own specified world to bring them up in and I’ll guarantee to take any one at random and train him to become any type of specialist I might select—doctor, lawyer, artist, merchant-chief, and yes, even beggarman and thief, regardless of his Martinez (2010) 15 talents, penchants, tendencies, abilities, vocations, and race of his ancestors. This bold proclamation leaves no question about where Watson stood on the nature/nurture question. Anyone interested in human behavior wants to understand how much of it is dictated by individual biology (nature) and how much is a consequence of the learning that occurs through life experiences (nurture). Watson took a very strong position on this question—he believed that any person could become anything given the right conditions. To John Watson, nurture was radically more potent than nature. Note an apparent paradox here: The strong role accorded the environment in shaping behavior speaks to human potential. Given the right conditions people could become essentially anything. At the same time, behaviorists denied that anybody has any choice in the matter. Such notions as wishes and intentions are excluded (most of the time) from the lexicon of the behaviorist. There is boundless potential, but no choice. This paradoxical duality (potential without choice) emerges even more explicitly in the theories of Watson’s successor, B. F. Skinner. Before considering Skinner’s particular contribution to behaviorism, operant conditioning, let’s be sure Martinez (2010) 16 we appreciate the significance of classical conditioning for education. Classical Conditioning and Education Classical conditioning shows how a neutral stimulus can evoke a simple behavioral response, including salivation (in the case of Pavlov’s dogs), or more significantly, strong emotional reactions, both positive and negative. Emotional reactions are important because they can strongly influence the choices we make and our perseverance in those choices. Using a bit of cognitive language here, emotions can shape our goals and aspirations, our beliefs about what is possible to achieve, and our convictions about what goals are worthy of our sustained efforts. Seen this way, the emotional associations forged by classical conditioning are highly significant for education. No one doubts that strong emotions are learned in educational settings. Classical conditioning can help us to understand how those emotions are learned. Much of the substance of education is relatively neutral in its ability to evoke emotional reactions. A mathematical equation is, by itself, neutral. But through experience, it can evoke interest and anticipation; alternatively, it can evoke anxiety, even fear. Likewise, the first page of a novel can prompt highly positive feelings of Martinez (2010) 17 anticipation or it can elicit feelings of dread. The world of education is filled with ideas, symbols, objects, and processes that have the potential to evoke strong emotions, which in turn can influence chains of decisions that accumulate to determine educational and career pathways. The course of a life can be channeled through sequences of such decisions. As educators, we want to encourage positive feelings about learning. Conversely, we want to avoid or reverse negative associations, especially dislike or anxiety. As we think about how emotions originate in learning situations, we should consider the roles of such negative stimuli as anger, shame, and criticism. We can also think of the positive associations of smiles, encouragement, and a friendly and affirming voice. Like Little Albert, each student has the potential to associate strong emotions with new experiences, people, and objects. The limitations of classical conditioning also must be acknowledged. For example, emotional reactions to specific academic subjects are more complexly determined than is suggested by their simple associations with pleasant or unpleasant experiences. They also involve beliefs about learning and about one’s own ability to learn. A more complete account of the emotional aspects of Martinez (2010) 18 learning seems to require cognitive language. Also—and this is an important limitation—classical conditioning explains only simple behavior. But we know that human behavior is complex. To understand and explain complex behavior, we need insights from a different form of behaviorism. We need to explore the theory of operant conditioning. E. L. Thorndike: Instrumental Conditioning Strictly speaking, E. L. Thorndike used the term instrumental conditioning rather than operant conditioning, but his research helped launch the operant conditioning paradigm as a new, and very powerful, form of behaviorism. Later, the basic principles of operant conditioning were expanded, sharpened, and popularized by the great behaviorist B. F. Skinner. But Thorndike’s contributions were a decisive turning point. Perhaps a little surprisingly, the foundations of operant conditioning were established early in Thorndike’s career, as part of his doctoral dissertation. Can the Cat Escape the Puzzle Box? Thorndike’s doctoral research employed a device that might seem a little cruel by today’s standards, but it gave Thorndike the information he wanted. The device was called a puzzle box, which was a Martinez (2010) 19 wooden cage in which Thorndike placed a cat. Thorndike was able to observe the cat’s behavior through spaces in the walls of the box. The key feature of the puzzle box was a small hatch that allowed the cat to escape. Unfortunately for the cat, the way to open it was not obvious. To open the hatch, the cat had to press an inconspicuous lever inside the cage. Thorndike wanted to know if a cat could escape from the puzzle box by pressing the lever, whether intentionally or accidentally. When placed inside the puzzle box, the cat exhibited random motions in frantic attempts to get out. Eventually, and without any particular strategy, the cat managed to hit the lever that opened the hatch. But the motion that activated the escape lever was accidental, a result of random behavior. The role of random behavior is important here, as it is in all operant conditioning. Random behavior can sometimes lead to results that are beneficial to the organism. Thorndike pioneered this concept, and later Skinner built an entire theory around it. Thorndike placed the cat inside the puzzle box repeatedly in order to understand how the cat adapted over time. Would the cat become more efficient in escaping from the puzzle box? Would subsequent escapes require less time? If so, how is that possible given that the initial escape was the Martinez (2010) 20 result of random behavior? Thorndike indeed found that the cat became more efficient with each subsequent trial. Each escape required less time. Learning was definitely taking place as the cat adapted to the puzzle box, but not immediately through insight into how to escape. The cat’s adaptation was gradual. This form of learning seemed unusual, but Thorndike had an explanation. The Law of Effect Thorndike knew that the adaptation of the cat to the puzzle box was the result of a gradual restriction of behavior. If the cat exhibited a wide range of random behavior initially, the scope of random behavior narrowed over time to actions that were more likely to be followed by a reward— specifically, escaping from the puzzle box. Perhaps only movements at one end of the box were associated with subsequent escape. Likewise, maybe only sideways movements of the paw were often rewarded by freedom. This is not to say that all of the repeated random movements were actually effective, but their association with a positive outcome (escape) was enough to encourage their repetition. Note that there is nothing mentalistic about learning here: The cat never had a singular moment of insight in which it knew what to do. Rather, the scope of random Martinez (2010) 21 behavior was gradually narrowed to behaviors that tended to be followed by a rewarding outcome. Based on these findings, Thorndike formulated the Law of Effect, which specifies that any behavior that is followed by rewards is more likely to be repeated than behavior that is not followed by rewards. The Law is commonsensical and intuitive; it might even seem obvious. But don’t forget what the Law explains in the case of the puzzle box; it shows how random activity, rather than insightful understanding, can lead to effective behavior. The Law of Effect can explain how a cat can learn to escape from a cage, yet it cannot explain the kind of learning that we are used to attributing to humans—learning that is conscious, insightful, and comprehending. But note a huge explanatory benefit of Thorndike’s Law of Effect: It can account for learning (conditioning) of complex behavior in a way that classical conditioning cannot. After all, the behavior of Thorndike’s cat was much more than a salivation reflex. Behaviorism now had a way of explaining how complex learning in animals could arise from the consequences of behavior. Soon, but in a different laboratory, the same basic paradigm would be applied to explain complex learning in human beings. B. F. Skinner: Operant Conditioning Martinez (2010) 22 It’s no exaggeration to say that B. F. Skinner was the greatest behaviorist of the twentieth century and that he turned the course of psychology in the United States. Even though behaviorism was built on the research and polemical arguments of Watson and Thorndike, Skinner himself was the true flagbearer of behaviorism for half a century. His legacy of influence extends beyond psychology theory to the popular culture; each of us has been affected by the ideas of B. F. Skinner, whether consciously or not. Figure 2.2 B. F. Skinner He changed the course of American psychology. [Copyright by the Archives of the History of American Psychology.] S-R-R: A Different Paradigm Skinner’s theory followed the same basic pattern of Thorndike’s: Animals and human beings could learn complex behavior gradually, as random actions were narrowed to be more and more effective over time. Skinner saw this kind of Martinez (2010) 23 learning as conditioning, but it was a different kind than the classical conditioning formulated by Pavlov. Skinner called his version of learning operant conditioning. Skinner (1938) formulated the theory of operant conditioning symbolically in a way that made it easy to compare with classical conditioning. Pavlov’s theory could be boiled down to two terms, stimulus and response, as follows: Stimulus → Response In Pavlov’s theory, the stimulus changes from the original, unconditioned stimulus to the conditioned stimulus. Still, only two terms were needed to make sense of learning in Pavlov’s laboratory. In Skinner’s account of operant conditioning, three terms were needed: stimulus, response, and reinforcement, as follows: Stimulus → Response → Reinforcement To clarify these terms, we’ll use the example of a dog learning to perform a trick, such as rolling over. The situation is simple: STIMULUS: The dog’s owner says, “Roll Over!” RESPONSE: The dog rolls over. Martinez (2010) 24 REINFORCEMENT: The dog gets a treat. Of course, when a dog learns a trick such as rolling over, the trick is not performed perfectly the very first time. It’s learned gradually, in a manner similar to the cat escaping from the puzzle box. Note another similarity between the cat escaping and the dog rolling over: What happens after the animal’s behavior is crucial. Whether we speak of Thorndike’s Law of Effect or Skinner’s concept of reinforcement, the consequences of performing a particular action are important. This is not the case in classical conditioning where behavior itself is not altered, only the stimulus needed to evoke a certain behavior. Classical conditioning involves learning, but it’s a different kind of learning than Thorndike and Skinner were studying. In Skinner’s theory, reinforcement is any consequence that leads to the repetition of a behavior. When a dog rolls over, reinforcement is provided in the form of a dog biscuit, or perhaps praising the dog and scratching behind its ears. For Skinner, any specific action that increases the likelihood of a behavior is a reinforcer. Reinforcers can be as basic as food or as abstract as money. Social reinforcers such as praise, smiles, and frowns can also be highly effective. Note that a reinforcer is defined by its effects: If a consequence increases the likelihood of a behavior Martinez (2010) 25 being repeated, then by definition it is a reinforcer. There is a subjective element here: What counts as a reinforcer cannot be specified completely beforehand. It will vary from person to person, and from one animal to another. What works as a reinforcer for one dog will not necessarily work for another; the same is true for people. Shaping Through Successive Approximations When a person or an animal learns a new “trick” through operant conditioning, the new behavior is typically learned gradually (Skinner, 1951). At first, the behavior only approximates the ultimate pattern in a coarse way. Rather than rolling over, the dog might at first lower its body closer to the ground. Over time, the dog might accidentally lower its body so that it touches the ground. Because this approximates the intended behavior more closely than before, the trainer of the dog might reward only this new behavior—touching the ground—with a treat. Now, in order to get a treat, merely lowering the body is not good enough; the dog must actually lower its body all the way to the ground. Eventually, this process can progress toward the target behavior of rolling over. As the animal more closely approximates the complex skill, the criteria for rewarding the dog become stricter. When the Martinez (2010) 26 dog rolls over for the first time, a new standard has been set. Now it is rewarded by praise or a biscuit only if it completes the trick successfully. This kind of progression is known as successive approximations. The terminology makes perfect sense: Over time, behavior evolves gradually toward successively greater approximations of its ideal form. A related term, shaping, describes the actions of a trainer or teacher to evoke more effective behavior over time. Behavior is shaped, much as clay is shaped by a sculptor. The behavior is not shaped instantly, but through successive approximations. Shaping is used to train animals to perform complex tricks to the delight of animal trainers and their audiences. Many kinds of animals—dogs, birds, elephants, sea lions, and killer whales, among others—can learn to perform complex behavior on command. Shaping has in fact been practiced since ancient times to domesticate animals for hauling, herding, and hunting (Timberlake, 1999). Figure 2.3 Shamu wows the crowd Operant conditioning can produce complex behavior in humans and animals. Martinez (2010) 27 [Photo courtesy of SeaWorld.] Interest Magnet 2.1: Teaching Shamu Principles of behaviorism apply not only to humans, but also to animals (Skinner, 1951). A visit to SeaWorld will give you an impressive demonstration of animal training using operant conditioning techniques. Through the skillful application of reinforcement by animal trainers, the orca or killer whale named Shamu has learned to leap through hoops, balance objects, and carry performers on his back. Some of the animal’s spontaneous behavior—a small leap from the water, for example—resembles what the animal trainer has in mind. If this act is reinforced, such as with a snack of fresh fish, Shamu will be more likely to repeat the small leap in the future. Gradually, the trainer raises expectations—Shamu receives rewards only when he leaps to greater heights than in the past. This method is called shaping or successive approximations. In stages, the rewarded behavior begins to resemble the impressive feats that customers want to see. Not only orcas, but also birds, elephants, dogs and many other kinds of animals can be trained through successive approximations. Martinez (2010) 28 How does Shamu learn to leap through hoops and toss human riders high into the air? Through patient application of the techniques of shaping. The whale learns amazing tricks by successive approximations. After being shown a hand signal, Shamu will at times approximate the desired behavior more closely, and at other times less closely. The human trainer carries a bucket of raw fish and judiciously rewards Shamu as his behavior approximates more closely the intended trick. It’s important to reinforce behavior immediately. That’s why trainers signal desired behavior with a whistle—called a “bridge signal” or secondary reinforcer—which promises that the whale will receive reinforcement in just a few seconds. Desirable behavior is reinforced by the bridge signal; undesirable behavior is not. That’s how learning progresses toward a specific behavioral goal. Stimuli other than hand signals are also used to indicate the desired behavior (SeaWorld, 2002). At SeaWorld, an orca learns to recognize computergenerated sound codes played on a waterproof keyboard. Other sound codes can indicate the name of the orca, the desired behavior, and even how that behavior is to be performed. The sound codes can also serve as bridge signals. Reinforcers other than fish are also used to promote specific Martinez (2010) 29 behaviors. These are advantageous because the whale might not be hungry during training. Other reinforcers include vigorous touching, squirting with a water hose, or supplying ice to eat and toys to play with. As each animal responds somewhat differently to various behavioral consequences, trainers must learn what works best for each animal. Training can contribute directly to an animal’s health. Shamu can hold still for obtaining blood samples or open his mouth for dental examinations (SeaWorld, 1994). He can even be trained to urinate for analysis on command. These “tricks” help the veterinarians at SeaWorld to monitor the animal’s health. Of course, Shamu saves his most impressive tricks for charming the audiences that watch him every day. Let’s return for a moment to Skinner’s term, operant conditioning. We know that conditioning is another word for learning in behaviorist terminology. But what does operant mean? An operant is a behavior that is generated by the animal, and as I’ve noted, it has a random element. Remember the cat in Thorndike’s puzzle box. Within the scope of a cat’s possible behaviors, it expressed or emitted a wide range of activity in its attempt to escape. Only a fraction of those random behaviors, or operants, had any real Martinez (2010) 30 effect. Thus, the scope of random operants tapered down over time as the cat was conditioned. The same random production of behavior, followed by the reinforcement of a subset of that behavior, can also account for a dog learning to roll over or a killer whale learning to jump through a hoop. Therefore, within Skinner’s theory there is a definite role for random behavior as the raw material for eventual complex behavior. Now consider parallels between two seemingly distant theories, Skinner’s theory of operant conditioning and Darwin’s theory of natural selection. Both recognize the importance of random variation as the raw material for changes in organisms. For Darwin, that random variation was genetic. Some subset of random genetic traits are beneficial to organisms as they adapt to their environments. Those traits, and their associated genes, are passed on to subsequent generations. In Skinner’s theory, random variation in behavior is displayed. Some of that random behavior—only a subset—is well-suited to the organism’s environment and tends to be repeated. Darwin showed how the selection of genes can change organisms from one generation to the next; Skinner showed how the selection of behaviors through reinforcement can shape an organism’s subsequent behavior. Normally, the conditioned Martinez (2010) 31 behavior is not passed on to subsequent generations. There is one major exception to this pattern—human beings. People excel at propagating knowledge from one generation to the next, and education plays a very prominent role in supporting this process. The parallels between operant conditioning and natural selection hint at the explanatory power of Skinner’s theory. After all, if Darwin’s theory is so powerful within biology, wouldn’t it make sense that a theory with a similar structure could also be quite effective in accounting for human behavior, and for explaining how behavior changes over time? Skinner certainly believed so. Like Darwin, Skinner thought his theory was capable of explaining all the phenomena under its purview. To him, the mechanisms of operant conditioning, such as shaping through successive approximations, are capable of accounting for all behavior. Operant conditioning is far more capable of explaining complex behavior than classical conditioning is. But can it really explain all complex behavior? Think of everything that you did today— all the thousands of skilled actions, including reading, writing, note-taking, making phone calls, banking, driving, exercising, dressing, eating, or shopping. Human activity is extremely complex even if we consider only directly observable Martinez (2010) 32 behavior and momentarily set aside (as Skinner would prefer) such cognitive concepts as thinking, planning, goal setting, and problem solving. According to Skinner’s theory, all behavior— everything that you did today, and in fact all your behavior through your entire life—is explainable within the theoretical system of operant conditioning. In other words, all that we do is the product of our reinforcement histories. Some of our past behavior was reinforced, some not, and this long chain of behavior and its consequences produced who we are and what we do today. Once we appreciate this sweeping claim, we understand that operant conditioning was intended to be a totalizing theory—one that purports to explain everything within its scope of interest. Teaching Complex Behavior If operant conditioning has the potential to be a totalizing theory to explain human behavior, why not use it also to control human behavior? That question certainly did occur to Skinner, who saw no reason not to use behavioristic principles to deliberately control human behavior. If applied carefully and systematically, reinforcement could shape the behavior of individuals in desirable directions. You might suspect, as Skinner did, that this was the basic goal of education anyway. Why not be deliberate about it and use operant Martinez (2010) 33 conditioning to be still more effective at controlling what people do? Behavioral modification is the name for any systematic approach to teaching new behavior using the principles of operant conditioning. We have already noted some of the key ideas in behavior modification, including shaping through successive approximations. We also know that reinforcement is important. For Shamu, it’s raw fish; for a dog, it’s a doggie biscuit. For human beings, money usually works well, but so do dozens of other rewards. Reinforcers, as we know, are defined strictly in terms of their effects—a reinforcer is any consequence to behavior that makes that behavior more likely to recur in the future. What works varies between people and across species. Shamu is not interested in cash, a dog does not want raw fish, and your next-door neighbor would not be motivated by a doggie biscuit. Punishment, like reinforcement, is also a significant factor in operant conditioning. If some consequences make particular behavior more likely, then other consequences make certain behavior less likely. If a police officer stops you for speeding, aren’t you less likely to drive fast in the future—at least in that vicinity? The theory of operant conditioning recognizes the importance of Martinez (2010) 34 punishment in learning. Reinforcement and punishment are complementary. One is a green light to continue the behavior; the other is a red light to signal that the behavior must stop. Reinforcement makes the recurrence of behavior more likely; punishment has the opposite effect. Punishers, like reinforcers, are always defined by their effects. Now that we have distinguished between reinforcement and punishment, another dimension to each must be recognized. We can speak of positive and negative reinforcement, as well as positive and negative punishment. Here, the terms positive and negative do not refer to whether an experience is pleasant or unpleasant. Instead, the terms denote whether a stimulus is being added or removed. Positive reinforcement is easy: it’s the doggie biscuit a dog gets for rolling over, or the large tip a waiter gets for providing good service. Negative reinforcement involves removing an aversive stimulus. When you get into a car, most likely an annoying ding ding ding will sound until you latch your seatbelt. Clicking the seatbelt buckle is reinforced by the removal of a stimulus— the irritating sound. More significantly, a prison inmate may have a prison term reduced for good behavior. That, too, is negative reinforcement. But be careful—in everyday speech, the term negative Martinez (2010) 35 reinforcement is often used incorrectly to refer to punishment. Punishment also has positive and negative varieties. Positive punishment is the application of an experience that is typically unpleasant: corporal punishment, an angry voice, or a threatening expression. They are positive, remember, because a stimulus is added. But punishment can also be accomplished by taking something away. Examples of negative punishment include the removal of privileges. When a child is prohibited from playing video games, watching television, or spending time with friends, these restrictions typically count as negative punishment. The widely applied techniques of “time out” for children’s bad behavior or “grounding” for teenage irresponsibility are also forms of negative punishment. Reinforcement Schedules The distinctions made in the previous section— between reinforcement and punishment, and between positive and negative varieties of each— are refinements in understanding how operant conditioning works. But Skinner and other behaviorists went much further in their analyses. They wanted to understand how the timing of reinforcement affects conditioning (Bijou, 1993). Martinez (2010) 36 For example, in a program of behavior modification, is it best to apply reinforcement after every instance of the desired behavior? Or should reinforcement be occasional rather than given every time? These questions bring us to the topic of reinforcement schedules. Suppose a father wants to apply behavioral modification to teach his daughter to make her bed every day. If the parent is highly motivated and has plenty of cash, he might pay his daughter a dollar each day she makes her bed (to some prescribed standard). When reinforcement is given every time the behavior occurs, the pattern of reinforcement is called continuous reinforcement. Continuous reinforcement is the most obvious reinforcement strategy; it’s very common, but it is not necessarily the most effective strategy. A father might wish that eventually his daughter will make her bed every day even in the absence of a cash reward; that is, one purpose of conditioning might be to reduce dependence on reinforcement over time. If that is the goal, then continuous reinforcement is not the best strategy to adopt. If continuous reinforcement is not the ideal strategy, what are the alternatives? There are four basic kinds of reinforcement schedules: Martinez (2010) 37 Fixed ratio Variable ratio Fixed interval Variable interval Fixed ratio schedules involve one reinforcement for every fixed number of displays of behavior. Suppose that father gives his daughter a dollar every other day that she makes her bed—or every third day. He can choose the number. If he is consistent in counting the instances of behavior and rewarding the behavior according to some ratio set beforehand—2 to 1, 3 to 1, and so on— then the reinforcement schedule follows a fixed ratio schedule. It is fixed because the chosen ratio does not change. Variable ratio schedules are different in that they involve a random element. Rather than reinforcing behavior after a predictable number of occurrences, reinforcement is applied after a randomly changing number of instances. In the example of a daughter making her bed, she might be rewarded with a dollar after two days of making her bed, or after 4 instances, or 10, then back to 3. The number jumps around and is unpredictable. The daughter knows that she will be rewarded only after she makes her bed, but never knows when that will occur. We will see that this reinforcement Martinez (2010) 38 schedule, variable ratio, has unusual power for motivating behavior and maintaining habits. Both fixed interval and variable interval schedules have a time factor, as in time interval. With fixed interval reinforcement schedules, reinforcement occurs after a predetermined period of time. A weekly paycheck fits this category. If the worker shows up for work and performs effectively, there will be paycheck ready to reward that worker at a predetermined time. With variable interval schedules, time is also a factor but the time interval between reinforcement is not fixed. Think about end-of-year pay bonuses. In some companies they are given out only in December, but bonuses might or might not be administered depending on the company’s profits. Similarly, quarterly stock dividends reward investors every three months, but they are not guaranteed. After disappointing fiscal quarters the company might not issue dividends to stockholders. Knowledge of these four reinforcement schedules can be very useful. But above all, we need to appreciate the significance of variable ratio reinforcement. This schedule is more powerful than the others. To understand why, think of what normally happens when a particular behavior is not rewarded. If you go to work every day but are not Martinez (2010) 39 paid, eventually you’ll consider not showing up. Similarly, a dog will expect a doggie biscuit now and then if he is to keep performing his tricks. If a month of rolling over produces no treat, he too might take a day off. Without reinforcement, conditioned behavior undergoes extinction. The behavior will diminish in regularity and intensity to some pre-conditioning baseline level, or it may stop altogether. No more dog tricks. There may be times when we want to instill behaviors in others—people or animals—such that the behavior does not require continuous reinforcement. What kind of reinforcement schedule makes a learned behavior most resistant to extinction? The variable ratio schedule. Figure 2.4 The unpredictable rewards of fishing Variable ratio reinforcement schedules can be highly compelling. [Photo courtesy of Duane L. Jellison, Registered Master Maine Guide.] Martinez (2010) 40 If a child learns that occasionally she will be rewarded for making her bed, but cannot predict when that reward will occur, she is most likely to maintain her new behavior for a long time—even in the absence of rewards. There is something about the unpredictability of rewards that makes the activity highly compelling. The power of variable ratio reinforcement is clearly seen in gambling. A gambler never knows when he will hit the jackpot. Even though there may be long runs of bad luck, his fortunes can change in a single card hand or in one roll of the dice. The power of variable ratio jackpots can help explain why gambling behavior persists despite intervals without reinforcement. The ongoing possibility of an imminent reward makes the behavior quite robust, for better or worse. Think about fishing, and you’ll see that the reward structure follows much the same pattern— behavior persists through long periods with no tangible payoff. Variable ratio reinforcement can work well when combined with the initial use of continuous reinforcement. That’s because continuous reinforcement is very effective for establishing a new behavior. Once the behavior is established, switching to a variable ratio reinforcement schedule can help make the behavior robust—that is, resistant to extinction in the absence of Martinez (2010) 41 rewards. Teaming the early use of continuous reinforcement with later use of variable ratio reinforcement is an effective strategy for behavioral modification. It’s a contribution from operant conditioning theory that has practical applications. Behaviorism’s Enduring Impact This book is not really about behaviorism. It focuses more directly on powerful concepts related to human cognition—thinking, reasoning, problem solving, memory, knowledge, beliefs, and many others. Nevertheless, behaviorism offers some really practical strategies, as we have seen. We must also appreciate the profound effects of behaviorism on our culture, especially on how we relate to each other on a human level. Part of living in a culture means that, collectively and as individuals, we try to influence if not control the behavior of other people, perhaps especially children. In our many and varied attempts to influence others’ behavior, Skinner’s version of behaviorism has had an impact that endures to this day. For example, we see behaviorism’s affect on education and parenting techniques. Although not discussed here, behaviorism also strongly influenced the theory and practice of psychotherapy. Still more broadly, the impact of Martinez (2010) 42 behaviorism was felt in social philosophy, in the way we think about society and culture. Impact on Education The strategic use of positive reinforcement is ubiquitous in education. Whenever elementary school teachers draw smiley faces on homework, positive reinforcement is used to encourage the continuation of the behavior. One smiley face or sticker might encourage the child to continue doing homework well and on time. Still more widely used is praise. To smile and say, “Good job, Ryan,” is to utter more than a spontaneous commendation. Praise is used to encourage the same behavior in the future. This is not to say that encouraging and positive comments to students are insincere. No doubt they are honestly intended. But teachers intuitively recognize the practical functions of praise—that smiles and commendations can help shape the future behavior of students. Indeed, teachers know that praise is one form of positive reinforcement that can be used to influence children’s behavior. Positive reinforcement is the most direct application of operant conditioning to teaching. Figure 2.5 Stars, stickers, and smiley faces Martinez (2010) 43 Teachers use positive reinforcers to shape the future behavior of students. [Photo by Jamie Stein.] Teachers use other principles of operant conditioning besides positive reinforcement. They deliberately withhold positive reinforcement in certain situations. When one student displays bad behavior, a teacher might advise the other students, “Just ignore him.” Why? Because there is a tacit awareness—an unspoken theory—that the child craves the attention that bad behavior typically provokes. The attention is the positive reinforcement. Without reinforcement the behavior will presumably taper off and undergo extinction. The hope is that bad behavior will be extinguished over time, not through active intervention but, counterintuitively, through deliberate in action. To ignore misbehavior rather than to confront it directly and forcefully is a break from past traditions of dealing with children when punishment was seen as the obvious response. Skinner himself opposed punishment, partly because he believed that it was not as effective as using, or withholding, reinforcement. The shift toward using reinforcement instead of punishment is one of the lasting contributions of operant Martinez (2010) 44 conditioning theory, though that fact is seldom recognized. The use and withholding of positive reinforcement are the most common applications of behavioral modification techniques in education, but there are others. Some have faded in popularity over time. In the heyday of behaviorism, Skinner (1958) applied operant conditioning to teaching techniques. The applications extended to so-called teaching machines, which were primitive precursors to computer-based learning systems. The instructional strategy was to break down the desired learning goals into discrete units of student behavior. Those units were arranged in an orderly sequence from easiest to most difficult. The teaching machine presented new information and called for responses to ensure that the student was learning. For example, the machine might present a new mathematical procedure followed by a practice problem. The student would write an answer on a roll of paper displayed within a frame. Having given an answer, the student would advance the written response so that it could not be altered. Next, the student would view the correct answer and so received immediate feedback. If the correct answer matched the student’s response, that match served as positive Martinez (2010) 45 reinforcement. The teaching machine would then advance to a more difficult problem. Figure 2.6 Skinner’s teaching machine The applications of operant conditioning to education were direct. [Copyright by the Archives of the History of American Psychology.] Although Skinner’s teaching machines have long become obsolete, some of the logic of teaching machines has continued to the present. Curricular practices of breaking down learning goals into small units calling for explicit student responses, and then giving feedback on those responses, has continued to the present. This way of organizing instruction has sometimes been transported to a different platform: a modern digital computer. Notwithstanding the rising sophistication of the machine used, behaviorism has had an enduring legacy in this approach. Whether delivered by machine or not, some forms of instruction break down material (and expected student responses) into very specific and discrete units. The Martinez (2010) 46 curriculum follows a highly prescribed logic that teachers are expected to follow closely. This form of teaching is known as programmed instruction. Even when instructional sequences are not so strictly prescribed, teachers are sometimes asked to identify goals of learning strictly in terms of what students will be able to do. When the goals of learning are expressed solely in terms of objective behavior, the resulting behavioral objectives are also a legacy of operant conditioning. Yet another influential movement in education was inspired by behavioristic principles—the approach called mastery learning. Promoted by Benjamin Bloom and his collaborators, mastery learning openly acknowledged that students learn at different rates. Mastery learning accommodated these different rates by breaking down the content of teaching into prescribed blocks and allowing students flexibility in the time needed to master a block of instruction. Mastery was operationalized by performance at some pre-established level, such as 80 percent correct on end-of-unit tests. Students would not be allowed to move on to a more advanced topic before mastering the prior unit. Elements of these varied behavioristic approaches are evident in current educational practices, even Martinez (2010) 47 though they are not as prominent as they were in the 1960s, 1970s, and 1980s. There is one segment of education, however, in which behaviorist techniques still hold considerable sway. Behavioristic principles continue to be employed broadly in teaching students with developmental delays or behavioral disorders. Effective teaching approaches among these special populations include highly prescribed systems of reinforcement. Sometimes, such tokens as poker chips are given as reinforcement. In this token economy, the chips can be exchanged later for toys or treats that are more directly appreciated by students. Impact on Parenting Behaviorism has had enduring effects on parenting practices, just as it has on education. Many of the reinforcement techniques adopted by teachers were appropriated by parents. Praising children in order to encourage the repetition of desired behavior has become a normal part of parenting practice. Likewise, many parents deliberately withhold attention for bad behavior—presumably to promote extinction of that behavior for lack of reinforcement. We see a version of this technique in “time out,” when children are forced to withdraw from play or social activity for misbehaving. Time Martinez (2010) 48 out can be seen as the withholding of reinforcement or as negative punishment. To be clear, the terminology of operant conditioning is not woven into the everyday conversations of parents. Perhaps in years past, parents thought about and spoke of positive reinforcement and conditioning more frequently than they do now. But even without the behavorist terminology, the use of operant conditioning techniques in current parenting practices affirms that Skinnerian ideas have entered the popular culture. Through theories of operant conditioning, Skinner and other behaviorists have had a lasting influence on the ways in which we try to modify the behavior of other people. Impact on Social Philosophy The principles of operant conditioning are so general and powerful that they are highly relevant to questions about how to regulate behavior in societies and cultures. Philosophers have long speculated about how a civil society is possible, and about the role of government in maintaining the social order and peaceful coexistence. Thomas Hobbes (1651/1963), for example, argued that one legitimate role of government is to stem the rabidly selfish motives of the average human Martinez (2010) 49 being. Restraint, in Hobbes’s theory, is a necessary function of government. Skinner’s theory, by contrast, does not assume that human beings have a greedy or selfish nature. Nor does Skinner’s theory assume that people are altruistic. Human behavior is neither basically good nor evil; it is simply the result of past reinforcement or punishment patterns in each person’s experience. Random behaviors (operants), allied with the selective force of consequences (reinforcement and punishment), can explain the entire range of human behavior. As John Watson claimed, experience alone is sufficient to transform any newborn infant into a saint or a sociopath. Reinforcement inevitably shapes the behavior of individuals over time. The pattern applies also to groups of people. Skinner knew that patterns of reinforcement and punishment run through a society or culture, and that these patterns can collectively shift the behavior of those societies. The effects are not always good. Societies may unwittingly reward aggressive acts or punish altruism. Indeed, counterproductive patterns of reinforcement must exist; otherwise we would not see violent crime in our society. All behavior, including all destructive behavior, can be fully explained by the patterns of reinforcement and Martinez (2010) 50 punishment experienced during each person’s unique history. Within the confines of operant conditioning theory, there is no other explanation. Strangely, the implied determinism of behaviorism removes blame from the criminal. What choice did he have, really—especially given that choice itself is not a reality according to the theory? Choice is cognitive talk, which behaviorism rejects. By seeing all present behavior as determined by the consequences of previous behavior, there is no true moral responsibility attendant to self-destructive, antisocial, violent, or traitorous acts. In an odd way, behaviorism is understanding and compassionate in its outlook. The proper response to undesired behavior is to reconfigure the reward structures. A society must replace one form of behavioral modification with another. If bad behavior is not blamed, neither is good behavior credited. If Jack the Ripper had no choice in the matter, neither did Mother Teresa. Good acts cannot be helped because, once again, behavior is strictly determined by our individual reinforcement histories. Freedom is therefore an illusion. So is virtue, integrity, and dignity. That is why B. F. Skinner (1971) could write a book entitled Beyond Freedom and Dignity. Both terms ascribe to humans choice guided by internal values Martinez (2010) 51 that have moral aspects. To Skinner, this was the wrong way to look at things. Ultimately, the illusions conveyed by such concepts as freedom and dignity, blame and credit, or virtue and vice impede progress toward a better social existence. Skinner did envision a utopian society in which peaceful coexistence reigned. That society, presented as fiction in Skinner’s 1948 novel Walden Two, was controlled by an intelligently structured logic of reinforcement. The “right” behaviors were consistently reinforced, producing a community that was cooperative, functional, and peaceful—not to say free, however. Perhaps like any version of a utopian society, the overarching system of behavioral modification running through Walden Two is chilling. The reader is struck by the fictitious society’s free use of manipulation, the robotic quality of its inhabitants, and the disturbing questions the novel raises about who establishes, maintains, and modifies the reinforcement contingencies that control the behavior of everyone else. Skinner’s version of an ideal society did not attract much serious interest. Perhaps the notion of a world run according to the principles of operant conditioning was not credible to anyone but the most ardent proponents of behaviorism. The social implications of behaviorism were Martinez (2010) 52 nothing more than a sidelight of Skinner’s theory, rather than a centerpiece. Nevertheless, consistency demanded that Skinner assert the implications of his theory for the behavior of entire societies. After all, if freedom is an illusion, then why not control all the contingencies of behavior rather than leave them to chance? In the end, though, maybe it was too much for anyone but academics and intellectuals to believe that freedom and choice are illusions. For the average person, it’s hard to accept that we can do without such concepts as plans, beliefs, purposes, goals, wishes, values, and intentions. Can behaviorism really explain everything about human behavior, or are other explanations needed? In time, the totalizing claims of operant conditioning theory were challenged, opening the way for alternative accounts of human behavior and learning. The alternative theories included explanations that acknowledged the role of the human mind, and explored its nature. Behaviorism Challenged As the influence of behaviorism grew during the twentieth century, its limitations gradually became apparent. Some of the problems of behaviorism were conceptual and definitional; others were related to inconsistencies between the tenets of behaviorism (in particular, the exclusion of Martinez (2010) 53 cognition) and what could be observed or inferred from the behavior of animals and humans. Eventually, criticism mounted to the extent that other theoretical approaches, including cognitive approaches, began to rival behaviorism for the dominant position in American psychology. Circular Definition of Reinforcement One conceptual problem with behaviorism is the definition of its central concept, reinforcement. As noted earlier, reinforcement is defined strictly in terms of its effects: namely, reinforcement is any consequence of behavior that increases the likelihood that the behavior will be repeated in the future. Defining reinforcement in this way makes some sense because what counts as reinforcement for one person (or animal) will not work for another. But there is also a problem created by this definition: It is impossible to test the idea that reinforcement causes behavior to be repeated. The two entities—repeated behavior and reinforcement—are linked by definition. Thus, behaviorism’s central declaration that reinforcement leads to the repetition of behavior is not testable. It is not a conclusion; rather, it is an assumption—an assumption, by the way, that can accommodate any evidence because it is not datadependent. Science relies on the testability of Martinez (2010) 54 assertions, and the possibility that propositions are falsifiable (capable of being disproved). In the scientific method, ideas are tested against data. Ideas may be vindicated or they may be vitiated; either way, science moves on. The definitional quality of reinforcement does not meet this criterion. Neither does the definition of punishment. Of course, science also relies on assumptions. The question is whether one can accept the definitions of reinforcement and punishment as assumptions rather than as testable claims. For some scholars, acceptance of such crucial circular definitions is asking too much. If so, this counts as one drawback of the operant conditioning form of behaviorism. Edward Tolman: Rats with Purpose One research-based finding by a prominent behaviorist proved to be subtly undermining to the overall theory. Experiments in the laboratory of Edward Tolman (1932) suggested that the behavior of rats could not be explained fully by behaviorist principles of stimulus and response. Tolman’s research involved studying how rats learn to navigate mazes. The rats in Tolman’s laboratory at the University of California, Berkeley, searched for food placed at the end of a path inside the maze. When rats ran the same maze in succession, they became highly efficient. But the rats could Martinez (2010) 55 also adapt whenever the maze was blocked by choosing an alternative, but still efficient, route. After observing thousands of such maze-running trials, the behaviorist Tolman could not escape a strong impression that the rats navigated the mazes as if they had a “mental map” of the maze structure. Here we must appreciate that a mental map, or knowledge of any sort, is a cognitive concept, not a behaviorist one. So-called radical behaviorists, such as Skinner, would not speak of mental maps or of knowledge, but only of behavior (Ringen, 1999; Skinner, 1990). Tolman went further. Not only did rats seem to have a mental map of mazes, they also appeared to run the mazes with a sense of purpose. It was as if the rats had a goal, an idea that guided their actions. For this reason, Tolman’s ideas are sometimes called purposive behaviorism (Innis, 1999). But the term purposive behaviorism is an oxymoron: Such notions as goals and purposes were anathema to strict behaviorists, just as references to mental maps and knowledge were forbidden. If these entities existed at all, they were hidden, internal characteristics of the organism, not the sort of strictly observable behavior that Watson and Skinner insisted we confine our discussions to. As early as 1932, Tolman began to question the strict dogma of radical behaviorism Martinez (2010) 56 with the publication of his book, Purposive Behavior in Man and Animals. Figure 2.7 Running mazes Rats seem to have mental maps and a sense of purpose. [Courtesy of Cartoon Stock, www.cartoonstock.com] Eventually, a new variety of behaviorists began to emerge. These neobehaviorists could not accept the stark restriction of their science to observable behavior. Instead, they felt it was necessary to consider the enduring characteristics of the organism. These internal characteristics—whether knowledge, goals, or motivation—influence behavior jointly with the external qualities of the environment. This meant that any science of behavior could not be built around only stimulus and response. A third entity, the organism, had to be brought into the conceptual picture. In fact, the basic paradigm of neobehaviorism could be symbolized by placing the organism between stimulus and response, as follows: Martinez (2010) 57 Stimulus → Organism → Response The new formula showed that a stimulus does not directly produce a response. By inserting the organism between stimulus and response, the paradigm of neobehaviorists explicitly acknowledged the mediating role of the organism in understanding how a stimulus can lead to a behavioral response. The relevant characteristics of the organism included such enduring traits as drive and purpose, as well as the strength of stimulus-response associations, called habit strength, caused by variation in reinforcement histories (Hull, 1943; Rashotte & Amsel, 1999). The S-O-R paradigm thereby breached the forbidden territory of the internal characteristics of the organisms—characteristics that were not directly observable but had to be inferred from behavior. While neobehaviorists were behaviorists still, their conceptual breakthroughs helped open the way some decades later to the advancement of cognitive theories. Noam Chomsky: Language Acquisition Device In time, behaviorism was attacked from yet other quarters. One salvo came from linguistics, and in particular from the linguistic theory of Noam Chomsky. To set the stage, let’s recall the strong claims of behaviorism that traced back to John Martinez (2010) 58 Watson. In Watson’s declaration, “Give me a dozen healthy infants...,” a strong theoretical position was declared. The position was that all human qualities are learned; every instance of complex behavior is fully a consequence of experience. Skinner sharpened the theory by stating precisely what kinds of experience exert such profound effects: the patterns of reinforcement and punishment that followed the behaviors of the organism. A long track of consequences determines the behavior of every organism. For humans, such behavior includes spoken language. This is where the linguist Chomsky had a problem with behaviorism. Human beings have such striking facility with language that, to Chomsky, a different conclusion seemed necessary—that humans are biologically programmed to learn language. Especially when considering how rapidly children learn language, some predisposition to learn language must be at work. That’s because language acquisition is not merely a matter of learning words. Beyond word knowledge, spoken language requires that the speaker learn rules of language use, called syntax or grammar. Chomsky noted, for example, that speakers can easily transform statements into questions, or questions into statements. Such transformations are commonplace, but who is ever taught how to Martinez (2010) 59 make them? Virtually every speaker uses such transformative grammars without being taught them directly. Instead, speakers infer syntax from the patterns of everyday speech. Mere experience with language prepares humans to acquire not only word knowledge, but also a sense of the underlying logic or deep structure of language. Such abstract understandings are acquired without instruction even by very young children. How is this possible? For Chomsky, it was inconceivable that this ability arises only through the stimulusresponse-reinforcement mechanism that Skinner believed to be sufficient. A different conclusion seemed inevitable: Only by carrying such readiness for language in their brain structure, and ultimately through genetics as coded in DNA, could such amazing facility with language be displayed with consistency in the human species. The paradigm wars between Chomsky and Skinner became somewhat personal when Chomsky published a review of Skinner’s book Verbal Behavior. In the book review, Chomsky (1959) articulated his objections to Skinner’s account of “verbal behavior” by citing the contrary evidence described above. Chomsky’s strong criticism of behaviorism as an all-encompassing, totalizing theory had a major effect on the field of psychology. Some see Chomsky’s review of Verbal Martinez (2010) 60 Behavior as a watershed between the previously dominant American behaviorism and a new era of cognitive psychology. A new consensus was forming; behaviorism could not account for all that was interesting and important about human behavior. Chomsky’s claim that the brain must be pre-wired for language helped to launch a more sweeping movement within psychology. That movement, called nativism, recognized that the human brain has some innate knowledge—unlearned ways of understanding the world and acting upon it—a product of our DNA-coded biology rather than our experience. Of course, biology must be conjoined with the right sorts of experience for the species’ heritage of native knowledge to realize its potential. But complex behavior is not merely a product of simpler behavior followed by consequences. Some pre-programming, Chomsky insisted, must be at work. If nativism holds any degree of truth, then sweeping claims by radical behaviorists (e.g., John Watson) and by “blank slate” philosophical empiricists (e.g., John Locke) must be incorrect. Albert Bandura: Learning Through Imitation Still other theorists were questioning the central ideas of behaviorism. One questioned assumption Martinez (2010) 61 was whether it is always necessary to experience firsthand the consequences of behavior. Do we learn only from the consequences of our own actions, or can we learn by observing what happens to other people? To avoid touching a hot stovetop, must we first burn our fingers? A more convenient alternative would be to learn from the mistakes of other people and so avoid their pain. Even better, perhaps we can learn to imitate behaviors that lead to pleasant consequences. The human capacity for imitation was compelling to psychologist Albert Bandura. He believed that firsthand experiences were not always necessary for learning to occur. He understood that human beings are highly capable of observing the behavior of other people and imitating that behavior—and of not imitating behavior that leads to disagreeable consequences. If one person becomes sick after eating leftover soup, there is no reason for others to repeat this mistake. They simply avoid the soup. The power of learning through observation was a cornerstone in Albert Bandura’s social learning theory, which recognized the major role of observation and imitation in learning. As such, it challenged the strongly individualistic mechanisms of learning identified by Skinner and other operant conditioning theorists. Even more important, social Martinez (2010) 62 learning theory required no obvious behavior on the part of the learner. Neither did it require reinforcement and punishment. Somehow, learning can take place without any of the elements considered so important by behaviorists. Instead, learning occurred through social observation. For example, Bandura documented social learning in children when those children watched film recordings of other children playing with a doll violently or placidly. Children who watched violent interactions with the doll were more likely later to punch and kick a real doll. Among children, merely observing the behavior of other children was enough to influence behavior (Bandura, Ross, & Ross, 1961). Humans surpass other animals in their ability to learn through imitation. Even primates such as chimpanzees are generally quite poor at imitative learning (Premack & Premack, 1996). This is not to say that other species do not imitate—they certainly do. But human beings are better at it— more consistent in acquiring knowledge and skill from others, and much more flexible in the range of behaviors that they can imitate. Indeed, the human capacity to learn from the experiences of other people helps to explain the depth and complexity of human culture, which has no rival in the animal world. The efficiency of learning is Martinez (2010) 63 greatly magnified through social observation. Direct experience is also important, of course, and has an essential role in education. But if all knowledge had to be reconstructed through direct individualistic experience, human culture would probably be forgotten within a single generation. The imitative processes highlighted in social learning theory strongly suggest that children acquire knowledge through observation, and of course knowledge is not an acceptable theoretical concept in Skinner’s radical behaviorism. To speak of knowledge is to employ cognitive terminology. Some years later, Bandura added another cognitive concept, self-efficacy beliefs, to advance our understanding of human behavior. Selfefficacy refers to a person’s sense of his or her own capability—whether or not the person is able to perform specific actions successfully. With this theoretical contribution, Bandura spoke not only of knowledge but also of self-knowledge, which pushed the theory still deeper into cognitive territory. From several different directions, Skinner’s bold theory of operant conditioning was attacked as either incomplete or incorrect. Chomsky and other nativists declared that human beings had certain capabilities that were inborn rather than learned solely from experience. The exclusive claims that Martinez (2010) 64 all learning was a product of behavior and its consequences therefore had to be qualified. Another qualification came from Bandura’s social learning theory. Especially in humans, learning was not always a product of individual experience. Learning could instead arise from observing the actions of other people, along with the consequences of those actions. Moreover, people act in ways consistent with their beliefs about what they can do. And, as Tolman observed, humans, rats, and other species seemed to act with a sense of purpose. The mighty castle of behaviorism was being shaken down to its foundations. In the castle walls, tiny cracks widened into fissures. Is There a Downside to Behaviorism? Theoretical challenges to behaviorism questioned whether learning always occurs through the experiences of reinforcement or punishment. These challenges were directed toward the center of Skinner’s behaviorism—how learning actually occurs. But behaviorism was also challenged from yet another quarter—research findings on the practical consequences of operant conditioning. There was no question that behavioral modification worked. Behavior could be altered through the studious application of reinforcement and punishment according to pre-defined reinforcement schedules. But, at least in some cases, behavioral Martinez (2010) 65 modification led to unanticipated and undesirable consequences. Undermining Intrinsic Motivation Typically, positive reinforcement is used to teach new behaviors—actions that would not normally be performed in the absence of reinforcement. But what if those behaviors were already expressed? And what if—to use a non-behavioristic concept— people actually enjoyed the activity prior to conditioning? What would happen if reinforcement were applied under such conditions? These questions were addressed in a classic experiment by Mark Lepper and his colleagues (Lepper, Greene, & Nisbett, 1973). To test the idea, the researchers selected an activity that was enjoyable to almost every child: playing with “Magic Markers,” felt-tipped pens that draw in a variety of bright colors. The research began by measuring how much time children played with the markers during free-choice periods. Preschool children needed no obvious reinforcement to motivate their play with the markers. They enjoyed playing with Magic Markers for their own sake. The children were intrinsically motivated. After the baseline levels of drawing were measured, children were randomly assigned to one Martinez (2010) 66 of three conditions. Children in the “expected award” group were promised a “Good Player Award,” a badge with ribbons showing the child’s name, if they agreed to make drawings for a visitor (the researcher) for about six minutes. These children exhibited interest in the Good Player Award, and each drew with markers to earn the badge. Children in the other two groups were not shown the award, nor were they told that they could earn a badge by playing with the markers. Children in the “unexpected award” condition were later given Good Player Awards even though they had not been shown or promised one beforehand. Children in the third, “no award” group did not know about the Good Player Awards, nor were they ever given one. Next, in the crucial part of the experiment, the researchers measured how much the children played with the Magic Markers after they were no longer associated with a promised reward. One week after children made the drawings, the researchers again observed the children during free-choice activities. Magic Markers and paper were made available to the children; the children could draw with the markers or not, as they preferred. The researchers observed the children through one-way glass so that their play times with the markers could be observed precisely, but Martinez (2010) 67 the children could not see the researchers. The researchers found that children in both the “no award” and “unexpected award” conditions did not change their overall level of play with the Magic Markers. Their use of the markers neither increased nor decreased from baseline levels. Figure 2.8 Children and Magic Markers If you reward children for drawing, will their intrinsic motivation be undermined? [Pearson Education/PH College] The key finding is that children who had been promised a Good Player Award, and given one, decreased their free play with Magic Markers. No longer motivated by the promise of an extrinsic reward, their voluntary level of play dropped below the baseline levels, and to about half the free-time levels of children in the other two groups. Having already earned a Good Player Award, the children were objectively less motivated to play with Magic Markers than they were before the experiment. Children in the second group—those never promised a badge but who were given one Martinez (2010) 68 anyway—did not differ in their baseline levels of free play with the Magic Markers. Finally, the quality of pictures drawn by the children during the experiments differed among the three groups. Drawings by children in the “expected award” condition were independently judged to be of lower quality than drawings made by children in the other two conditions. How are these findings to be interpreted? Lepper and his colleagues inferred that the promise of a reward had decreased the intrinsic motivation that children in the first group originally had for playing with Magic Markers. Initially motivated to play with Magic Markers because of the inherent enjoyment they gave, children reinforced to play with the Magic Markers no longer felt as intrinsically motivated. Extrinsic rewards—the Good Player badges—had undermined intrinsic interest. As a result, voluntary play with the markers dropped off, and the drawings instrumental to earning the award were poorer in quality than drawings made by children who did not expect an award. The Negative Effects of Extrinsic Rewards Around the same time as the Magic Marker experiment, other studies confirmed that extrinsic rewards could undermine initial intrinsic interest in an activity. For example, when adults were paid Martinez (2010) 69 money for solving a puzzle, their free-time engagement in the puzzle decreased (Deci, 1971). Even token amounts of money were enough to undermine intrinsic interest. This body of research suggests a warning: Extrinsic rewards have the potential to undermine intrinsic motivation. This conclusion, stated as a possibility rather than a certainty, is highly relevant to education for two reasons. First, operant conditioning is strongly based on the use of extrinsic rewards. Positive reinforcement, as presented by Skinner, is the application of reinforcers that are not directly connected to the intended behavior. The two are distinct, connected only by the reinforcement contingencies set up by the teacher, parent, or animal trainer. Any reinforcer, such as food, money, or a badge, might decrease preexisting intrinsic motivation for an activity. A second reason to be wary of the use of extrinsic rewards is that intrinsic motivation is universally recognized as an important learning outcome. Virtually every teacher is concerned not simply with teaching particular behaviors, but also is dedicated to increasing students’ intrinsic motivation for learning. Intrinsic motivation has the wonderful benefit of sustaining learning long after any particular course has come to an end. Martinez (2010) 70 Tangible rewards are much less common in schools than is praise. Is there a worry, then, that the use of praise as a positive reinforcer might undermine intrinsic motivation, just as a badge or money could? The answer is less clear. Among students with lower academic ability, the connection between course grades and praise appears to be positive; however, some research has revealed a slight negative correlation between praise and academic achievement among highability students (Brophy, 1981). It seems possible that praise could undermine the intrinsic motivation of higher-ability students—but this is speculation. The data only suggest this possibility. Nevertheless, caution is in order: Any extrinsic reward, even praise, has the potential to reduce students’ initial intrinsic motivation for an activity, including learning. An undermining effect is not certain, but teachers should keep this possibility in mind when they use reinforcement techniques. Learning Strategies This chapter presented an approach to learning that has a distinct theoretical basis: Behaviorism concentrates on observable behavior and how to change the behavior. The two forms of behaviorism, classical and operant conditioning, both have practical applications to teaching and learning. Some of these applications, such as the Martinez (2010) 71 use of positive reinforcement, have found their way into traditional teaching practices. Beyond positive reinforcement, behaviorism offers several other possible applications to education. This section presents ten strategies to promote learning that build directly on the theory presented in the preceding pages. 1.Avoid associating learning with negative emotions. Research on classical conditioning shows that neutral objects can evoke strong negative emotions, such as fear. Unfortunately, many students feel negative emotions when dealing with school subjects, books, classrooms, and teachers. Teachers can minimize these negative emotions by trying to avoid inciting feelings of fear, anxiety, or shame among students. Negative emotions may effectively motivate student behavior in the short term, but have counterproductive consequences in the long term. Whenever negative emotions are evoked by school contexts, the result is a very poor context for advancing the goals of learning. Martinez (2010) 72 2.Associate learning with positive emotions. The counterpart to avoiding conditioning of negative emotions is to intentionally associate positive emotions with learning. When a mathematics teacher evokes laughter and smiles, and generates a sense of emotional safety among students, those feelings can become associated, subtly but significantly, with the subject of study. More than one student has been positively influenced by a teacher’s affirming personality. Such influences may have been significantly a matter of classical conditioning—the association of positive feelings with neutral objects. The idea is basic, but it is powerful, nonetheless. 3.Use continuous reinforcement to initiate new behavior. Classical conditioning works well for teaching simple forms of behavior, but for complex learning it’s necessary to rely on operant conditioning. One principle of operant conditioning is that to establish a new behavior, the most effective approach is to reinforce that behavior every time it Martinez (2010) 73 occurs. Remember, the organism is inclined to repeat only behaviors that are reinforced. If reinforcement is not applied consistently in the beginning, the child (or the animal) might abandon the newly learned behavior. 4.Tailor reinforcement to the individual. Skinner defined reinforcement only by its effects—a reinforcer is whatever causes a behavior to be repeated in the future. That same principle applies to differences among people; a reinforcer must be tailored to the preferences of individual students. A hug might be reinforcing to one student but punishing to another. Likewise, public praise might be strongly rewarding to one student, but a major disincentive to another. This complicates the work of a teacher or parent, certainly. But to believe that there is a set category of reinforcers that works for every person every time is simply wrong, and can well be counterproductive. 5.Be careful not to reinforce undesired behavior. Operant conditioning theorists were well aware that the anger evoked by Martinez (2010) 74 misbehavior might well be reinforcing to the miscreant student. When a student misbehaves and the teacher’s attention shifts immediately to the student (“Stop doing that!”), the student might find the focused attention very rewarding. The teacher thereby mistakes reinforcement (what the student feels) for punishment (what the teacher intends). This simple error is one that most teachers understand very well. We can thank behaviorists for explaining why what seems superficially to be punishment is actually a reward. 6.Reduce reinforcement after the new behavior becomes established. After a new behavior has been established, a teacher can gradually reduce reinforcement over time. As reinforcement is gradually withdrawn, the new behavior can become self-sustaining, motivated by its own intrinsic rewards. Extrinsic rewards can establish the new competency; but once established, intrinsic rewards can sustain it indefinitely. The progression toward intrinsic motivation for learning is a worthy goal in the value system of every teacher. Martinez (2010) 75 7.Switch to random reinforcement to make the new behavior robust. Continuous reinforcement is effective in establishing a new behavior, such as paying attention to schoolwork. But once a behavior is established with some regularity, the teacher can employ a new strategy. Now, the goal is to strengthen the new behavior so that it does not have to be reinforced every time. For example, a teacher not only wants each learner to pay attention to the task at hand, but also to develop the habit of paying attention. The goal is to make the new behavior resistant to extinction so that it will persevere through long periods without reinforcement. To strengthen the new behavior, it’s a good idea to switch to a variable ratio schedule of reinforcement— to gradually reduce the frequency of reinforcement and to introduce an element of random timing. Variable ratio schedules can help establish the behaviors as habitual and enduring. 8.Be careful of the undermining effects of extrinsic rewards. Martinez (2010) 76 Following the widespread use of behavior modification in education and childrearing, psychologists found a potentially serious drawback to the use of extrinsic awards—they might undermine intrinsic motivation. Whenever using extrinsic rewards, a teacher should bear in mind this possibility: If a student has an abiding interest in a topic, that interest could actually decrease as a consequence of extrinsic rewards. Course grades, ubiquitous in schools, may well have such undermining effects. Other extrinsic motivators may add to this effect. The undermining potential of reinforcement is not necessarily inevitable—rewards might help to establish a nonexistent behavior. Still, all teachers and parents should be aware of this possible pitfall of operant conditioning. 9.Use unanticipated rewards. In some experiments on extrinsic rewards, unanticipated awards do not undermine intrinsic motivation. For example, when extrinsic rewards are given as unanticipated surprises for wanted behavior they can help sustain that behavior without subtracting from Martinez (2010) 77 students’ intrinsic interest (Lepper et al., 1973). Skillful teaching might include weaving in unanticipated rewards so that intrinsic enjoyment of the activity can continue and, ideally, build up over time. 10. Use cognitive concepts. For all the conceptual and practical benefits of behaviorism, the theory has serious limitations for guiding teaching and learning. Any complete account of teaching practice must include such cognitive concepts as reasoning, problem solving, critical thinking, interest, curiosity, and understanding. These concepts are not part of the vocabulary of behaviorism, which concentrates exclusively on observable behavior. Teachers need not—indeed, should not— restrict themselves conceptually in this way. That is why when teachers apply behavioristic principles, it’s best not to rely on behaviorism exclusively. Rather than treating behaviorism as a sufficient theory for understanding and influencing all human behavior, a teacher is better advised to be theoretically eclectic, mixing and matching ideas from various theories Martinez (2010) 78 as they open up possibilities for advancing the goals of education. These ten strategies show how the principles of behaviorism—both classical and operant conditioning—can advance the goals of teaching and learning. The applications are straightforward and practical, but they are not exhaustive. Like all the principles presented in this book, the ten listed are strategies rather than foolproof rules. Every strategy must be adapted to the characteristics of particular learners and to the unique approach and personality of the teacher. Not all strategies need to be used—perhaps only two or three will work well for any particular teacher and situation. When those strategies are used, they should be applied sensitively and sensibly to the particular mix of learners who, with their teacher, are jointly advancing what their minds can do. Conclusion This chapter surveys the historical and conceptual high points of behaviorism, from the physiological experiments of Pavlov to the social philosophy of Skinner. In its range of concepts and applications, behaviorism—both classical and operant conditioning—can account for a wide range of human and animal learning. But the kinds of learning accounted for by classical and operant Martinez (2010) 79 conditioning are different. Classical conditioning can explain how initially neutral stimuli can evoke relatively primitive behavior, including fear reactions. Operant conditioning helps to explain how complex behavior can be learned through a succession of reinforcement and punishment consequences. Looking back on the history of classical and operant conditioning, we should remember what motivated the restriction of behaviorism to observable phenomena. John Watson articulated the stance most clearly: He believed that psychology could evolve as a scientific field most expeditiously if it confined its range of relevant phenomena to objective behavior. Because such entities as knowledge, goals, beliefs, and intentions are not directly observable, they cannot be reliably studied—or so behaviorists believed. Eventually, the self-imposed delimitations of behaviorism paid off with theories that illuminated conditions of learning. As we have seen, behaviorism—particularly the principles of operant conditioning—has affected teaching and childrearing practices in significant and enduring ways. Above all, the legacy of behaviorism is manifest in the deliberate and strategic use of positive reinforcement to encourage the repetition of desired behavior. Positive reinforcement is used Martinez (2010) 80 routinely by teachers and parents. More obscurely, the use of certain schedules of reinforcement, especially the variable ratio schedule, can make learned behavior robustly resistant to extinction in the absence of rewards. For all these conceptual and strategic benefits, however, behaviorism does have certain disadvantages. In its application, the use of positive reinforcement can, under some circumstances, undermine pre-existing intrinsic motivation. If students expect their engagement in an activity to be followed by a reward, the students might conclude that they engage in that activity only because of the reward—not because they enjoy the activity. This possible effect of positive reinforcement should suggest caution to every teacher and parent who applies techniques of behavioral modification. But there is a more significant downside to strict reliance on behaviorism as a way to understand and influence learning. The language of behaviorism, especially in its original and more radical formulations, excludes the language of cognition. The exclusion of cognitive concepts from discussions about learning is a huge liability and disadvantage for anyone interested in education. Educators need to consider not only reinforcement contingencies, but also such vital concepts as Martinez (2010) 81 reasoning, understanding, mental models, problem solving, and critical thinking. Education is enriched by taking such concepts seriously because teachers understand that students have minds, and that students’ minds must learn to reason and understand as part of the education process. Moreover, all learners have interests and motives, hopes and fears. A complete theory of learning, as well as a complete theory of effective teaching, needs such powerful ideas at its center. We can speculate whether or not behaviorism was good for American psychology. Some have argued that its conceptual restrictiveness over a half-century cost psychology dearly (Norman, 1990). Only in the 1960s did cognitive psychology germinate and then flourish among psychologists. Inspired by the metaphor of a digital computer, rigorous models of the human mind began to shed light on the nature of the thinking process, including problem solving. The theoretical payoff was tremendous. We now know much about how the mind works, the nature of knowledge, and the many varieties of complex cognition. Not only has our knowledge of the mind advanced, but we also know better than ever how to promote understanding and complex thought. Our era is perhaps the best yet for teaching and teachers. Today’s teachers can profitably apply the legacy of Martinez (2010) 82 principles derived from behaviorism; but those principles, however useful, are more powerfully complemented by discoveries that emerged from the study of human cognition. References Bandura, A., Ross, D., & Ross, S. A. (1961). Transmission of aggression through imitation of aggressive models. Journal of Abnormal and Social Psychology, 63, 575–582. Bijou, S. W. (1993). Behavior analysis of child development (2nd rev.). Reno, NV: Context Press. Brophy, J. (1981). Teacher praise: A functional analysis. Review of Educational Research, 51(1), 5–32. Chomsky, N. (1959). A review of B. F. Skinner’s “Verbal Behavior.” Language, 35(1), 26–58. Deci, E. L. (1971). Effects of externally mediated rewards on intrinsic motivation. Journal of Personality and Social Psychology, 18(1), 105– 115. Hobbes, T. (1651/1963). Leviathan: On the matter, forme, and power of a commonwealth ecclesiasticall and civil. New York: Collier Books. Martinez (2010) 83 Hull, C. (1943). Principles of behavior. New York: Appleton-Century-Crofts. Innis, N. K. (1999). Edward C. Tolman’s purposive behaviorism. In W. O’Donohue & R. Kitchener (Eds.), Handbook of behaviorism (pp. 97–117). San Diego, CA: Academic Press. Lepper, M. R., Greene, D., & Nisbett, R. E. (1973). Undermining children’s intrinsic interest with extrinsic reward: A test of the “overjustification” hypothesis. Journal of Personality and Social Psychology, 28, 129–137. Norman, D. A. (1990, August 26). The mind exists: Commentary on the death of B. F. Skinner. Los Angeles Times (Op-ed page). Pavlov, I. P. (1927). Conditioned reflexes. London: Routledge and Kegan Paul. Premack, D., & Premack, A. J. (1996). Why animals lack pedagogy and some cultures have more of it than others. In D. R. Olson & N. Torrance (Eds.), The handbook of education and human development (pp. 302–323). Oxford: Blackwell. Rashotte, M. E., & Amsel, A. (1999). Clark L. Hull’s behaviorism. In W. O’Donohue & R. Martinez (2010) 84 Kitchener (Eds.), Handbook of behaviorism (pp. 119–158). San Diego, CA: Academic Press. Ringen, J. (1999). Radical behaviorism: B. F. Skinner’s philosophy of science. In W. O’Donohue & R. Kitchener (Eds.), Handbook of behaviorism (pp. 159–178). San Diego, CA: Academic Press. SeaWorld (2002). SeaWorld/Busch Gardens Animal Information Database. www.seaworld.org. SeaWorld (1994). Training Shamu. U.S. Kids. Skinner, B. F. (1938). The behavior of organisms. New York: Appleton-Century-Crofts. Skinner, B. F. (1948). Walden two. New York: Macmillan. Skinner, B. F. (1951). How to teach animals. Scientific American, 185(12), 26–29. Skinner, B. F. (1958). Teaching machines. Science, 128, 3330, 969–977. Skinner, B. F. (1971). Beyond freedom and dignity. New York:Knopf. Skinner, B. F. (1990). Can psychology be a science of mind? American Psychologist, 45, 1206– 1210. Martinez (2010) 85 Timberlake, W. (1999). Biological behaviorism. In W. O’Donohue & R. Kitchener (Eds.), Handbook of behaviorism (pp. 243–284). San Diego, CA: Academic Press. Tolman, E. C. (1932). Purposive behavior in animals and men. New York: Century. Watson, J. B. (1913). Psychology as the behaviorist sees it. Psychological Review, 20, 157– 177. Watson, J. B. (1924). Behaviorism. New York:Norton. eBook Collection Home Table of Contents Read Book Online Buy This Book Download eBook Print Chapter Search Preferences Help Useful Links Reference and Citation Generator Martinez (2010) APA Information Grammar and Writing Guides Home | Classroom | Library | Program | Account Sitemap | Downloads | Community Relations | System Status | Terms of Use | Privacy Policy Contact Us Report a problem | Ask a question | Share a thought Copyright © 2010 University of Phoenix PWAXDNET007 86